Remote chiral induction in the organocatalytic hydrosilylation of aromatic ketones and ketimines

Article abstract of DOI:10.1002/anie.200503941

(Chemical Equation Presented) Lewis basic, metal-free pyridyloxazolines catalyze the reduction of prochiral aromatic ketones and ketimines with Cl ₃SiH in good enantioselectivity (up to 94% ee). Arene-arene interactions between the substrate and the catalyst are likely to play a role in the enantiodifferentiation process.

Full text of DOI:10.1002/anie.200503941

Communications
Asymmetric Reduction
DOI: 10.1002/anie.200503941
Remote Chiral Induction in the Organocatalytic
Hydrosilylation of Aromatic Ketones and
Ketimines**
Andrei V. Malkov,* Angus J. P. Stewart Liddon,
Pedro Ramírez-López, Lada Bendovµ, David Haigh,
ˇ
´
and Pavel Kocovsky*
Dedicated to Professor Henri Kagan
on the occasion of his 75th birthday
Asymmetric hydrogenation, hydroboration, and hydrosilyla-
tion are the most frequently used catalytic methods for the
reduction of prochiral ketones 1 and imines 2 (Scheme 1).[1]
Scheme 1. Catalytic reduction of ketones and ketimines; for R1–R3, see
Tables 1 and 2.
[*] Dr. A. V. Malkov, A. J. P. S. Liddon, Dr. P. Ramírez-López, L. Bendovµ,
ˇ
´
Prof. Dr. P. Kocovsky
Department of Chemistry, WestChem
Joseph Black Building
University of Glasgow, Glasgow G128QQ (UK)
Fax: (+44)141-33-488
E-mail: amalkov@chem.gla.ac.uk
p.kocovsky@chem.gla.ac.uk
Dr. D. Haigh
Department of Medicinal Chemistry
Metabolic and Viral Centre of Excellence for Drug Discovery
GlaxoSmithKline Pharmaceuticals
Medicines Research Centre
Stevenage, Hertfordshire, SG12NY, (UK)
[**] We thank the GSK and EPSRC for an industrial CASE awards to
A.J.P.S.L., the Ministry of Education and Science of Spain for a
postdoctoral fellowship to P.R.-L., the Socrates Exchange program,
and Dr. Alfred Bader and the University of Glasgow for an additional
support.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
1432
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1432 –1435

Angewandte
Chemie
Although asymmetric hydrogenation remains favored by
industry in general, it is not free of problems, namely, those
associated with metal leaching, high pressure, and the cost of
the catalyst and its regeneration. Stoichiometric borane
reduction, catalyzed by chiral oxazaborolidine, avoids most
of these problems and offers high levels of enantioselection,[1]
but its cost is prohibitive for large-scale industrial application.
The recently developed reduction of imines, which uses the
Hantsch dihydropyridine as a stoichiometric reducing agent
and a chiral Brønsted acid as an organocatalyst,[2] is also
tainted by the cost implications. Transition-metal-catalyzed
hydrosilylation[1,3,4] on the other hand relies on much cheaper
silanes as reducing agents but shares the problem of metal
leaching with hydrogenation. Furthermore, the current meth-
ods usually perform well with either ketones or imines, but
rarely with both classes.[4] The alternative metal-free meth-
ods[2,5] are rare and considerably less effective;[6] nevertheless,
a promising development has recently been reported,[7] which
relies on Cl3SiH, an inexpensive and easy-to-handle reducing
agent.[8] Herein, we report an organocatalytic hydrosilylation
applicable to both ketones and ketimines.
oxazoline 6, obtained from (S)-mandelic acid, proved much
more reactive and furnished 3a in 78% ee (entry 2).[10]
Reduction of a series of aromatic ketones 1b–e, catalyzed
by 6, proceeded in a similar fashion, with 70–80% ee
(entries 3–6). On the other hand, reduction of the non-
aromatic ketone 1 f resulted in the formation of a racemic
product (entry 7). Other solvents, such as toluene, diethyl
ether, and THF, proved inferior.
The lower reactivity of 5 relative to 6 can be conjectured
to originate from an increase in the steric radius of the extra-
coordinate silicon moiety by coordination to the ligand. Thus,
the coordination of Cl3SiH to 5 is impaired by the adjacent
phenyl group (Scheme 2, A), whereas this adverse effect is
absent in chelate B, derived from 6.[11]
As part of our program focusing on the activation of
organosilicon reagents by Lewis bases,[9] we examined 2-
pyridyloxazolines of type 5, derived from (S)-phenylglycinol,
as potential organocatalysts in the hydrosilylation of aromatic
ketones with Cl3SiH (Scheme 1, 1!3). Initial attempts to
reduce acetophenone (1a) in the presence of 5 (20 mol%,
CHCl3, ꢀ208C, 24 h) resulted in a rather low conversion into
alcohol 3a (Table 1, entry 1). By contrast, the isomeric
Scheme 2. Coordination of Cl3SiH to (2-pyridyl)oxazolines.
To improve the efficacy of the catalyst and to shed more
light on the mechanism, related oxazolines were prepared
with the quinoline 7 or isoquinoline 8 and 9 fragments in place
of the pyridine moiety (Scheme 1). No reduction was
observed with 7 (CHCl3, ꢀ208C; Table 1, entry 8), presum-
ably because of the steric constraints imposed by the
quinoline moiety, thus mirroring the behavior of 5. The 3-
isoquinoline catalyst 8 was approximately as efficient as the
parent (pyridyl)oxazoline 6 (entry 9; compare with entry 2).
A real improvement over a range of substrates (1a–e,g,h) was
attained with the 1-isoquinolyl catalyst 9 (entries 10–16),
which allowed the catalyst loading to be lowered to 10 mol%
and provided by far the best enantioselectivities in the metal-
free hydrosilylation of ketones to date (ꢁ 94% ee).
Table 1: Reduction of ketones (1!3) with trichlorosilane catalyzed by
5–9.[a]
Entry Catalyst (mol%) Ketone R1, R2
Yield [%] ee [%][b,c]
1
2
3
(S)-5 (20)
(S)-6 (20)
(S)-6 (20)
1a
1a
1b
Ph, Me
Ph, Me
2-MeO-C6H4,
Me
29[d]
85[d]
66[g]
78
77
100[d]
4
5
6
7
8
9
10
11
(S)-6 (20)
(S)-6 (20)
(S)-6 (20)
(S)-6 (20)
(R)-7 (20)
(R)-8 (20)
(S)-9 (10)
(S)-9 (10)
1c
1d
1e
1 f
1a
1a
1a
1b
2-F-C6H4, Me
30[d]
70
80
80
0
Importantly, catalyst 9 was also found to exhibit high
efficiency in the reduction of N-aryl imines 2a–f (Table 2).
Therefore, it can be regarded as the first organocatalyst that
4-Me-C6H4, Me 100[d]
Ph, Et
c-C6H11, Me
Ph, Me
Ph, Me
Ph, Me
91[d]
67[d,e]
0
–
41[e,f]
85[f]
50[f]
73[g]
84
87
Table 2: Reduction of ketimines (2–4) with trichlorosilane catalyzed by
9.[a]
2-MeO-C6H4,
Me
Entry
Imine
R1, R2, R3
Yield [%][b]
ee [%][c,d]
12
13
14
15
16
(S)-9 (10)
(S)-9 (10)
(S)-9 (10)
(S)-9 (10)
(S)-9 (10)
1c
1d
1e
1g
1h
2-F-C6H4, Me
4-Me-C6H4, Me
Ph, Et
2-naphth, Me
6-Me-2-naphth,
Me
35[f]
90[f]
55[f]
93[f]
93[f]
70
85
86
94
92
1
2
3
4
5
6
2a
2b
2c
2d
2e
2 f
Ph, Me, Ph
65
60
67
67
51
65
87
Ph, Me, PMP[e]
85
2-naphth, Me, Ph
2-naphth, Me, PMP[e]
4-MeO-C6H4, Me, PMP[e]
4-CF3-C6H4, Me, PMP[e]
87 (98[f])
86
87
87
[a] The reactions were carried out on a 0.4-mmol scale with Cl3SiH
(2.1 equiv) in CHCl3 in the presence of the catalyst at ꢀ208C for 24 h.
[b] Determined by chiral GC or HPLC. [c] All products 3 were R config-
ured (unless stated otherwise), as revealed by comparison of their
optical rotation with the literature data. [d] Conversion determined by GC
of the reaction mixture after a standard work up. [e] The reaction was
carried out in CH2Cl2. [f] Yield of the isolated product (shown to be pure
[a] The reactions were carried out on a 0.4-mmol scale with HSiCl3
(2.0 equiv) in CHCl3 in the presence of catalyst 9 (20 mol%) at ꢀ208C for
24 h. [b] Yield of the isolated product (product shown to be pure by
1H NMRspectroscopic analysis). [c] Amines 4 were R configured, as
revealed by comparison of their optical rotation and their HPLC retention
times with the literature data. [d] Enantiomeric excess was determined by
HPLC. [e] PMP=p-methoxyphenyl. [f] After a single recrystallization
from methanol.
1
by H NMRspectroscopic analysis). [g] The product was S configured.
naphth=napthyl.
Angew. Chem. Int. Ed. 2006, 45, 1432 –1435ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1433

Communications
allows reduction of both ketones and ketimines with a
practically useful enantioselectivity.
To rationalize the high selectivity of catalysts 6, 8, and 9,
the following issues have to be taken into account: 1) Table 1
reveals that the order of reactivity of the ketones displays a
good correlation with their Brønsted basicity[12] (pKHB): 2-
MeO-C6H4COMe
(1.34) > PhCOMe
(1.11) > 2-F-
C6H4COMe (0.90). If we accept that this trend can be
extrapolated to Lewis basicity, the results would suggest that
the carbonyl oxygen atom is coordinated to the weakly Lewis
acidic Cl3SiH in the transition state.[13] 2) Trichlorosilane does
not reduce carbonyl or heterocarbonyl compounds unless
activated by Lewis bases through the formation of extra-
coordinate silicon species, such as C, in the case of bidentate
activators (Scheme 3).[7] Coordination of the carbonyl group
Figure 1. Arene–arene interactions of acetophenone (1a) with pyridine
as the model TS E.
magnitude of this interaction lends credence to the proposed
involvement of p stacking in the transition state E.
A similar mechanism can be envisaged for the reduction
of imines. However, the coordination of Cl3SiH to the
nitrogen atom is unlikely, as this would create a very
crowded environment. Hence, the R3 group of imine 2 can
be assumed to fill the space that the SiHCl3 group fills in E.
Experimental data support this hypothesis: Note that the
electron-rich ketones react faster than their electron-poor
counterparts (Table 1, entries 3 versus 4 and 11 versus 12),
which is consistent with the carbonyl basicity (see below)
and, therefore, its propensity to coordinate the Lewis acidic
Cl3SiH. By contrast, imines exhibit the opposite trend
(Table 2, entries 5 versus 6).
In conclusion, we have developed a new, practical, metal-
free protocol for the enantioselective reduction of aromatic
ketones (ꢁ 94% ee) and ketimines (ꢁ 87% ee). The reaction
is characterized by an unusual, long-ranging chiral induction.
The enantiodifferentiation is presumed to be aided by
aromatic interactions between the catalyst and the substrate.
Scheme 3. Mechanism of hydrosilylation.
(D) might be presumed (upon replacement of a chloride atom
at the silicon center) but can be regarded as unproductive,
since the corresponding four-membered transition state (TS)
for the reduction is unlikely.[13,14] 3) A linear relationship
between the enantiopurity of catalyst 6 and the product was
observed for the reduction of acetophenone (1a!3a), thus
suggesting that only one molecule of the catalyst is actively
involved in the enantiodifferentiating event. 4) The role of the
substituent on the oxazoline rings in 6, 8, and 9 can be
attributed to the shielding of one of the faces of the catalyst by Experimental Section
the Ph group.[15]
General procedure for the asymmetric reduction of ketones 1 and
ketimines 2 with trichlorosilane: Trichlorosilane (86 mL, 0.84 mmol,
2.1 equiv) was slowly added dropwise to a solution of the catalyst
(11.0 mg, 0.04 mmol or 21.9 mg, 0.08 mmol) and the corresponding
ketone or imine (0.40 mmol, 1.0 equiv) in CHCl3 (2 mL) at ꢀ208C.
The reaction mixture was stirred at ꢀ208C for 24 h, after which time
saturated aqueous NaHCO3 (1 mL) was added to quench the
reaction. The mixture was extracted with CH2Cl2 (3 ” 10 mL) and
the combined organic extracts were dried over MgSO4. Concentration
in vacuo followed by flash chromatography on silica gel (2 ” 15 cm)
with petroleum ether/ethyl acetate (20:1) or CH2Cl2 as the eluent
afforded alcohols 3a–3h or amines 4a–4 f.
Accordingly, the following mechanistic picture can be
drawn: the N,N-chelation of Cl3SiH by the catalyst creates an
activated hydrosilylating species, while another molecule of
Cl3SiH is likely to activate the ketone by coordination to the
oxygen atom in the E fashion.[16] The ketone–Cl3SiH complex
will then approach the catalyst–Cl3SiH complex from the less-
hindered side (as dictated by the remote chiral center in the
catalyst). Transition state E, which accommodates all these
effects, is consistent with the experimentally observed si-facial
selectivity of the reaction and with the features discussed
herein. It can also be hypothesized that E will be stabilized by
arene–arene interactions between the heteroaromatic sys-
tems of the catalyst and the substrate.[17]
Received: November 7, 2005
Published online: January 20, 2006
The potential role of the latter p–p interactions in shaping
the TS[17] was assessed by a computational analysis of the
pyridine–acetophenone model complex (Figure 1).[18] The
successive scans, according to the steering angle (q), the
distance between the planes (r1), and the parallel displace-
ment of the two molecules (r), were carried out (each scan
starting at the geometry of interaction energy minimum of the
previous scan). The lowest interaction energy found by the
potential-energy surface scans was at ꢀ6.3 kcalmolꢀ1. The
Keywords: arene–arene interactions · asymmetric catalysis ·
N ligands · organocatalysis · reduction
.
[1] E. N. Jacobsen, A. Pfaltz, H. Yamamoto, Comprehensive Asym-
metric Catalysis, Vol. I–III, Springer, Heidelberg, 1999.
[2] a) S. Singh, U. K. Batra, Indian J. Chem. 1989, 27, 1; b) M.
Rueping, E. Sugiono, C. Azap, T. Theissmann, M. Bolte, Org.
Lett. 2005, 7, 3781; c) S. Hoffmann, A. M. Seayad, B. List,
1434
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 1432 –1435

Angewandte
Chemie
Angew. Chem. 2005, 117, 7590; Angew. Chem. Int. Ed. 2005, 44,
7424.
[17] Arene–arene interactions have been suggested by us to account
for the stereochemical outcome of allyl silane addition to
aromatic aldehydes;[9d,e] Hayashi and co-workers arrived at a
similar conclusion: a) T. Shimada, A. Kina, T. Hayashi, J. Org.
Chem. 2003, 68, 6329; Birman et al. recently invoked arene –
arene interactions to rationalize the kinetic resolution in the
acylation of racemic alcohols: b) V. B. Birman, H. Jiang, Org.
Lett. 2005, 7, 3445.
[3] O. Riant, N. Mostefai, J. Courmarcel, Synthesis 2004, 2943.
[4] A highly efficient Cu-catalyzed hydrosilylation of ketones and
ketimines has been developed: a) B. H. Lipshutz, K. Noson, W.
Chrisman, A. Lower, J. Am. Chem. Soc. 2003, 125, 8779; b) B. H.
Lipshutz, H. Shimizu, Angew. Chem. 2004, 116, 2278; Angew.
Chem. Int. Ed. 2004, 43, 2228.
[18] The geometries of monomers were determined by the cc-pVTZ/
RI-MP2 gradient optimization in the Turbomole v.5.6 package
(R. Ahlrichs, M. Bꢀr, M. Hꢀser, H. Horn, C. Kölmel, Chem.
Phys. Lett. 1989, 162, 165); the optimized geometries of
monomers were kept frozen throughout the scans. Interaction
energies were computed with aug-cc-pVDZ basis set at RI-MP2
level in Turbomole v.5.6 and are BSSE-corrected. Only the face-
to-face geometry was examined, thus keeping the planes of the
monomers parallel throughout the scans.
[5] a) P. I. Dalko, L. Moisan, Angew. Chem. 2001, 113, 3840; Angew.
Chem. Int. Ed. 2001, 40, 3726; b) P. I. Dalko, L. Moisan, Angew.
Chem. 2004, 116, 5248; Angew. Chem. Int. Ed. 2004, 43, 5138;
c) E. R. Jarvo, S. J. Miller, Tetrahedron 2002, 58, 2481; d) A.
Berkessel, H. Gröger, Asymmetric Organocatalysis, Wiley-VCH,
Weinheim, 2005.
[6] a) R. Schiffers, H. B. Kagan, Synlett 1997, 1175; b) H. Nishikori,
R. Yoshihara, A. Hosomi, Synlett 2003, 561; c) F. J. LaRonde,
M. A. Brook, Tetrahedron Lett. 1999, 40, 3507; d) F. J. LaRonde,
M. A. Brook, Inorg. Chim. Acta 1999, 296, 208; e) M. D. Drew,
N. J. Lawrence, W. Watson, S. A. Bowles, Tetrahedron Lett. 1997,
38, 5857.
[7] a) R. A. Benkeser, D. Snyder, J. Organomet. Chem. 1982, 225,
107; b) S. Kobayashi, M. Yasuda, I. Hachiya, Chem. Lett. 1996,
407; c) M. A. Brook, Silicon in Organic, Organometallic and
Polymeric Chemistry, Wiley, New York, 2000, pp. 133 – 136; d) F.
Iwasaki, O. Onomura, K. Mishima, T. Maki, Y. Matsumura,
Tetrahedron Lett. 1999, 40, 7507; e) F. Iwasaki, O. Onomura, K.
Mishima, T. Kanematsu, T. Maki, Y. Matsumura, Tetrahedron
Lett. 2001, 42, 2525.
[8] A very efficient organocatalytic protocol for the hydrosilylation
of ketimines (with ꢁ 92% ee) has been developed by us
recently: a) A. V. Malkov, A. Mariani, K. MacDougall, P.
ˇ
´
ˇ
Kocovsky, Org. Lett. 2004, 6, 2253; b) A. V. Malkov, S. Stoncius,
ˇ
´
K. N. MacDougall, A. Mariani, G. D. McGeoch, P. Kocovsky,
Tetrahedron 2006, 62, 264.
[9] a) A. V. Malkov, M. Orsini, D. Pernazza, K. W. Muir, V. Langer,
ˇ
´
P. Meghani, P. Kocovsky, Org. Lett. 2002, 4, 1047; b) A. V.
Malkov, M. Bell, M. Orsini, D. Pernazza, A. Massa, P. Herrmann,
ˇ
´
P. Meghani, P. Kocovsky, J. Org. Chem. 2003, 68, 9659; c) A. V.
ˇ
´
Malkov, M. Bell, M. Vassieu, V. Bugatti, P. Kocovsky, J. Mol.
Catal. A 2003, 196, 179; d) A. V. Malkov, L. Dufkovµ, L.
ˇ
´
Farrugia, P. Kocovsky, Angew. Chem. 2003, 115, 3802; Angew.
Chem. Int. Ed. 2003, 42, 3674; e) A. V. Malkov, M. Bell, F.
ˇ
´
Castelluzzo, P. Kocovsky, Org. Lett. 2005, 7, 3219.
[10] Hydrosilylation of 1a, catalyzed by Rh complexes of 5 and 6,
produced alcohol 2a in 69% ee and 9% ee, respectively: H.
Brunner, U. Obermann, Chem. Ber. 1989, 122, 499.
[11] Similar effects have been observed by us for substituted
bipyridines, in which the increased bulk prevented coordination
of [Mo(CO)6] and other metal complexes: A. V. Malkov, I. R.
Baxendale, J. Fawcett, D. R. Russel, V. Langer, D. J. Mansfield,
ˇ
´
M. Valko, P. Kocovsky, Organometallics 2001, 20, 673.
[12] F. Besseau, M. Luçon, C. Laurence, M. Berthelot, J. Chem. Soc.
Perkin Trans. 2 1998, 101.
[13] T. Kudo, T. Higashide, S. Ikedate, and H. Yamataka, J. Org.
Chem. 2005, 70, 5157.
[14] Transfer hydrogenation and oxazaborolidine-catalyzed reduc-
tion of ketones with borane have been shown to proceed through
a six-membered transition state: a) M. Yamakawa, H. Ito, R.
Noyori, J. Am. Chem. Soc. 2000, 122, 1466; b) S. Itsuno in
Comprehensive Asymmetric Catalysis, Vol. 1, Springer, Heidel-
berg, 1999, pp. 290 – 315.
[15] Replacing the Ph in 6 with tBu led to virtually identical results,
which rules out arene–arene interactions of the substrate with
the Ph in 6.
[16] For a review on the carbonyl coordination (E vs Z; h1 vs h2), see:
S. Shambayati, W. E. Crowe, S. L. Schreiber, Angew. Chem.
1990, 102, 273; Angew. Chem. Int. Ed. Engl. 1990, 29, 256.
Angew. Chem. Int. Ed. 2006, 45, 1432 –1435ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1435